I think it's best to defer any thought of switching d-i to a newer kernel until after rc3. At the moment rc3 is nearly ready (except for some missing kernel abiname updates), and I don't see any reason why we cannot get it released within the month. Switching kernels is sure to take longer than that, and we'd run the risk of becoming a sarge freeze blocker, since rc2 is not considered suitable for the final release.
I do think we could consider 2.6.10 after rc3 is released, if the sarge release is still being indefinitly delayed by other factors at that point. Independant of concerns over which kernel sarge releases with, I think that it's beginning to be important to have at least an unstable version of d-i that uses the newer kernels. We've talked about ways to do this without destabalising the sarge d-i; it's hard, but not impossible. If we're going to do that anyway, then we can have a real basis for comparison of d-i with 2.6.8 vs d-i with 2.6.10, and if the latter proves to work well enough we can look at using it for a d-i release. If it doesn't, at least it will be available. It seems from the meeting and other discussion that there are no really release critical issues in the 2.6.8 kernel that affect d-i and cannot be at least worked around. Some of the issues raised should probably be added to the d-i errata -- could someone go back through the transcript and pick out errata items? As to 2.6.10's suitability for release, my impression from the meeting and other discussion is that we really don't know, and that it's likely to have RC items of its own, especially on less popular architectures, that will need more kernel updates to address. -- see shy jo
signature.asc
Description: Digital signature
